[Closed] Budget DIY Progammable 0-10 Volt Decade and 4.5 Digit Voltmeter

[quantumvolt]

Coming - need time to start up writing ...

Goal: 0-5 Volt (Optional 10.5 V) Programmable 5 Decade Voltage Reference.

   - Step 1 mV. 16 Bit DAC Monotonic output. Expected error less than 0.1%.

   - Open Buffered DAC Version Programmable from Keypad on Arduino.

   - Closed Loop DAC/ADC Version with Digital Feedback (Software / 20 Bit ADC).

   - 4.5 Digit Voltmeter (20 Bit ADC) on Arduino. Accuracy better than 0.1%.
      Code from Linear Technology Linduino LTSketchbook

   - DMM Voltage Checker / Calibrator 0.05% (max error). No calibration needed.
     Output voltages: 2.5, 5.0, 7.5, 9.096, 11.596 Volt DC.

Budget: USD 25-60 (parts list price) + an Arduino (clone) w. I2C Display 
    (USD 15-20 total) + 12 V wall wart + 3 pcs. 9 V batteries
    + general electronics bits, pieces & scrap ...



Construction base :





http://cds.linear.com/docs/en/demo-board-schematic/934asch.pdf

[quantumvolt]

My latest toys. One of the boards has Dual 16 Bit DAC and Dual 20 Bit ADC. The DVM Shield has 24 Bit ADC. The software window is the free QuikEval Demo Board Software from LT, but they do not give code for this program - just an .exe-file. So I will not use it much. However - the C++ code library LTSketchbook for their new Arduino-compatible Linduino board is free, so I will write all new snippets using that library. I am also tired of writing new GUI code for Arduino's Serial Window every time I start a new project with new data and data-format, so I have decided to get a licence for MakerPlot (the full screen program - USD 39).

I have never worked with a DAC before, so it will be interesting. Especially the Closed Loop DAC / ADC with Digital Feedback (Software - no hardware). I find that having to write is a good way to learn myself. I have reserved 3 posts that I will use to keep all information, links and videos in one place. Got the idea from some of the scope review test threads.

I am so tired of breadboards that I have decided to hack the Linear Technology Demo Board DC934A (USD 50 for the DAC / ADC board - list price around USD 25 for the chips only). Hopefully big fun  . I'll be back with more ...





Board:
http://cds.linear.com/docs/en/demo-board-manual/dc934af.pdf from
http://www.linear.com/demo/dc934a

MakerPlot: http://www.makerplot.com/

[Harvs]

Hey thanks for the tip on MakerPlot.  At first glance it looks quite like something I've been looking for, for some time now.  I ended up getting an academic licence of matlab, but makerplot looks like it could be really quick to get data up and running for quick streaming demos and the like.

As for the variable reference project, I'm doing something somewhat similar.  I need to be able to calibrate data acquisition channels on another project I've built (this other project I'm not the sole owner of the IP, so cant discuss it on here.)  So anyway I'm still waiting on parts to arrive from farnell, but basic outline is the same with the following parts:
- LTC2440 ADC
- LTC6655CHMS8-5 reference
- 2x AD5662 16bit DACs (will be configured as per the LT app note for higher res.)
- LTC6090 140V opamp
- ebay Vishay bulk foil resistor for the feedback divider (i.e. second hand pulls)
- Si RF isolators for serial
- parts for a 120V linear supply.

And I'll probably use a STM32 Cortex M0 as I have a few floating around.  However, doesn't really matter what uC for this app, as it's pretty slow.  I've also got some AVRs and PIC32s.

I've got a 1U 19" rack case I'm going to put it in, and through USB be able to auto cal the DAQ channels.  Needs to have a range of true zero to 100V.  As I'm calibrating 16bit DAQ channels, the resolution and accuracy should be fine.

[quantumvolt]

MakerPlot is nice - although a little slow. But I (and my arms and eyes) are tired of writing one-off junk programs in C++ and then a week later not any longer remember which of the versions is for what. If all you do is solder, write code and look at an 8 bit osc traces, then there is no metrology. So I go for the MakerPlot even if it is a little bit limiting. Nice software all in all ...

Your setup looks nice and in a class higher then what I was thinking of. My plan was to write the tutorial I needed but never found when I started to get interested in Volts. All I got was a few of the salesmen on the forum suggesting buying a ready made hobbyist voltage reference. So I did - and a single one has proved to be excellent. The others (in total more than USD 150 in purchases) have drifted out of spec within 3 months. For low / medium end references it is nothing but stupidity in my view to buy newly produced not yet sufficiently burned in refs that have been trimmed with a pot for calibration. A little bit dust and a new smell in your home and the ref is off by 50 ppm ...

So this project would have used 3 refs with initial accuracy 0.025 % and no trimming (hardware adjustment kills the history and 'age' of the gadget - software calibration leaves the device untouched and stable). By series and back-to-back on 3 batteries you can generate dispersed voltages from around 1.6 to 11.5 volts. Measure these combinations with your most stable handheld DMM (because that is all you have if you go for this project - people with 6.5 digit DMM's don't reflect on it). By a simple linear regression from a web applet or spreadsheet you get a best fit straight line and errors / correction constants for your DMM.

When done, the DAC and ADC is a piece of cake. The ADC is exactly what I did in the DVM thread, and the DAC is only different in that the output is I2C (as is the display) - not SPI (which it is for the ADC). No external components needed. Some wires, a few decoupling caps and some pull-up resistors etc. and you are plug-and-play in 1 or 2 factors of 10 better accuracy then your average good DMM.

Anyway - I have a lot of things to do. So I move on to the 18 (and later on 20) bit DAC. I now have high end ADC and DAC devices from LT, MicroChip and AD